Table of Contents Author Guidelines Submit a Manuscript
Applied Bionics and Biomechanics
Volume 2018, Article ID 6452050, 11 pages
https://doi.org/10.1155/2018/6452050
Review Article

Automated Techniques for the Interpretation of Fetal Abnormalities: A Review

1Department of Bio Medical Engineering, Samrat Ashok Technological Institute, Vidisha, India
2Department of Electronics and Instrumentation Engineering, Samrat Ashok Technological Institute, Vidisha, India
3Department of Electronics and Communication Engineering, G. B. Pant Government Engineering College, Delhi, India

Correspondence should be addressed to Vidhi Rawat; moc.liamffider@lraep_ihdiv

Received 22 December 2017; Revised 7 April 2018; Accepted 10 May 2018; Published 10 June 2018

Academic Editor: Stefano Zaffagnini

Copyright © 2018 Vidhi Rawat et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Abstract

Ultrasound (US) image segmentation methods, focusing on techniques developed for fetal biometric parameters and nuchal translucency, are briefly reviewed. Ultrasound medical images can easily identify the fetus using segmentation techniques and calculate fetal parameters. It can timely find the fetal abnormality so that necessary action can be taken by the pregnant woman. Firstly, a detailed literature has been offered on fetal biometric parameters and nuchal translucency to highlight the investigation approaches with a degree of validation in diverse clinical domains. Then, a categorization of the bibliographic assessment of recent research effort in the segmentation field of ultrasound 2D fetal images has been presented. The fetal images of high-risk pregnant women have been taken into the routine and continuous monitoring of fetal parameters. These parameters are used for detection of fetal weight, fetal growth, gestational age, and any possible abnormality detection.

1. Introduction

There are various types of imaging modalities available such as ultrasound system, CT scan, MRI, NMR, and X-rays. There are various display modes, but the brightness mode ultrasound is the most normally applied investigative tool due to its noninvasive nature, cheaper cost, and small risk to the patient compared to other image modalities [1]. Generally, in radiology, injections such as radio-opaque dyes are needed, but in US, imaging external source is not required [2]. For diagnosis, the images of the organ are the most powerful technique for the obstetrician and gynecologist [3]. US image is molded when the satisfactory beam of sound waves is sent through the transducer in the human body. Received echo by the replication from internal organs creates appropriate ultrasound images. Moreover, due to properties of image formation, they could be influenced by the speckle, attenuation, missing boundaries, and artifacts, making the segmentation assignment more complicated [4].

The National Consensus for Medical Abortion in India report specified due to complications related to abortion each year an average of about 11 million abortions occur annually and around 20,000 deaths due to complications related to abortion [5]. Precise fetal parameter dimensions of US images are key issues for the pregnant woman’s better health care. In obstetrics, fetal biometric parameters and thickness of nuchal translucency are essential parameters for the detection of fetal abnormality. The fetal biometric parameters include gestational sac (G.Sac), biparietal diameter (BPD), head circumference (HC), abdominal circumference (AC), and femur length (FL). These biometric parameters are used to measure the gestational age of the fetus and detect the growth patterns and abnormalities [1].

The nuchal translucency (NT) thickness of the fetus at 11–14 weeks of gestation was used to diagnose chromosomal abnormality [6]. NT thickness is the fluid accumulation in the nuchal region in the first trimester. Extensive research has verified that Down syndrome is a specific disorder triggered by the presence of an additional chromosome on chromosome 21. Generally, every human cell comprises 23 pairs of diverse chromosomes. Every chromosome transmits genes which are desirable for appropriate growth of human bodies. For the duration of conception, a specific receives 23 chromosomes each from the mother and the father. Children may receive the additional chromosome from any one of the parents. The latest study demonstrates that such fetal chromosomal anomalies can be sensed by measuring the NT thickness in the first trimester. The normal and abnormal growth is detected through measurements with the population-based growth chart. Manual measurements of fetal parameters are subjected to inter- and intraobserver variability [7]. Automatic methods for fetal parameter measurement reduce the inconsistency and create more accurate and reproducible measurements [8, 9]. Automated fetal monitoring improves the workflow efficiency; it helps to efficiently measure the fetal parameter. These accurate measured parameters will help the radiologist to diagnose the status of the fetus [10].

An examination of medical images including image acquisition, enhancement, segmentation, compression, and storage of the measurement of anatomical and physiological parameters is presented [11]. The segmentation process of images gives qualitative and quantitative image analysis. The weak edges and wrong edges are inherent in the US images. It is more problematic to correctly segment the images. Many reviews [1217] on image segmentation have been published in different journals, but none focused on the segmentation of 2D ultrasound fetal medical images. Figure 1 shows the process flow diagram for fetal growth detection.

Figure 1: Process flow diagram for fetal growth detection.

Three-dimensional medical ultrasonography was described in the early 1990s for fetal screening, but its spread was inadequate due to poor image class and slow acquisition protocols, unable to prevent fetal motion artifacts [18, 19]. These limitations are gradually vanishing with cutting-edge technologies, increasing the clinical interest in 3D ultrasound (3DUS). During the first trimester and early stage of the second trimester of gestation, the field of view of the ultrasound probes can integrate the whole gestational sac. Consequently, 3DUS-based volumetric studies of uterine structures have been published [20], as well as quantification of the whole fetus [21] or partial body portions (e.g., head and trunk) [22], providing useful information for clinical routine. These volumetric studies still rely on manual tracing, and automated segmentation methods are, therefore, desirable. Semiautomated methods were used in recent studies, especially with the software tool VOCAL, commercialized by General Electric and cited in several works [2123]. It enables to reconstruct smooth organ surfaces from a set of 2D contours acquired on rotated views along a single axis [24]. This software remains limited to the extraction of single organs and is not yet capable of segmenting complex objects such as the whole fetus. Moreover, several manual interactions are often needed. The cost of 3DUS is also very high, so generally radiologists prefer 2DUS. In this paper, we focus only on 2DUS.

This paper is organized as follows. Section 1 is the introduction of the automated techniques for the interpretation of fetal abnormalities; Section 2 describes the enhancement techniques of US images. Section 3 summarizes the segmentation techniques of measuring the fetal parameters. Section 4 presents the future trends of segmentation techniques. Finally, concluding remarks and suggestions for further development are outlined in Section 5.

2. Preprocessing of Ultrasound Images

Enhancements of US images are essential in manual assessment as well as computer-based analysis. US images are formed due to the pulse-echo system so the discrimination between normal and abnormal regions is complicated. The echoes received by the transducer are to be subject to the characteristic impedance of the medium: where and are the characteristic impedance of the medium, is the reflected ultrasound beam (echoes), and is the incident beam.

Procedures for contrast enhancement of ultrasound images are well known. The radiologist receives US images which contain arbitrary variations due to the statistics of echoes produced from the object. The detection of small and slight structure is difficult due to noises [25]. Speckle noises are formed from backscattered echoes randomly dispersed in the tissue [26, 27]. Because of the speckle presence, radiologists sometime fail to reach the conclusion [28]. The presence of speckle noise in the US images bounds its application in medical imaging. As a result, edge preservation [29] and enhancement [30, 31] are an essential operation in ultrasound image processing. The images are enhanced by applying various statistical filters [32], and the results are proven by measuring different parameters. where is the estimated local contrast. The indigenous contrast delivers high-frequency noise; is the image gray level and is the median gray level inside the region of . Eq. (2) can be equated to a high-pass spatial filter. A Bayesian estimator-based discriminator for the improvement of images by extrication of image and noise was proposed [33]. It is a semiblind noise removal algorithm founded on a steerable wavelet pyramid.

3. Extraction of Fetal Parameters by Segmentation Techniques

In obstetrics [3], the fetal biometric parameters and nuchal translucency are the key parameters to indicate any possible abnormalities in the fetus. The normal growth of the fetal body indicates the changes in shape across gestation weeks of the fetus.

3.1. Fetal Biometric Parameters

The US system is noninvasive in nature, so continuous fetus monitoring is safe to use in the obstetric field. Assessment of the growth of the fetus and diagnosis of the fetal abnormality is easy using the segmentation process in image processing. Mostly, the image investigation is based on 2-dimensional B-mode US images. Among all biometric fetal parameters, head and abdomen segmentation is simple because of texture similarities and clear boundaries. The femur of the fetus can lack internal texture which makes the extraction more difficult. Abdomen and whole fetus segmentation is harder due to inconsistencies in the internal structures. The fetal biometric parameter measurement methods are used limitedly in clinical practice [9, 3436].

3.1.1. Probabilistic Boosting Tree (PBT)

PBT classifiers are represented by the nodes of a binary tree. Binary classification of data sets is automatically clustered by PBT [37]. Carneiro et al. [34] automatically detected the fetal parameters by the segmentation process applied on US images. The fetal parameters were also measured by ultrasound images based on the development of a constrained probabilistic boosting tree. In this work, automatic measurement of BPD, HC, AC, and FL of the fetus has been presented. They patented and developed a marketable system, called auto-OB [9]. This system used in clinical practice is the only system for measuring the fetal parameters.

3.1.2. Fuzzy Logic

Fuzzy logic is an exceptional methodology applied across ultrasound images due to the fact that it does not require exact and enhanced images. In 1996, a semiautomatic fuzzy decision system developed for examination of the fetus has been presented. The system relies on the enhancement of the acquired images, which follows the decisional algorithms in the form of a sequence of If-Then rules. After acquiring the raw image and converting it into a desired format, various image processing algorithms are applied to analyze the images and measure the femur length, head circumference, and abdominal circumference [38]. Further, fetal biometric parameters are measured and analyzed by the maximum likelihood (ML) criterion algorithm, as proposed by Jardim and Figuiredo [39]. Manual extraction of contour in medical images requires expert knowledge and higher processing time. Fetal biometric parameters are measured for the detection of gestational age by a class-separable shape-sensitive approach [40]. In this approach, too many cost functions are assumed, which shows both its limitations and complications. The cost and objective function is the mathematical expression for the shape-sensitive derivative approach. The cost function at a different pixel level of the image is given by [41] where is the cost function and is the weighting factor. The value of depends on the intensity of images and the number of classes.

3.1.3. Thresholding-Based Morphological Operator

The femur of the fetus is segmented through the morphological operator, then the length of the femur is measured. Thomas et al. [42] proposed the morphological feature-based algorithm to detect the contour of the femur in US images and automatic length measurement of the femur bone in the fetus. Further, in 2009 [43], gestational age of the fetus was measured using a femur length. Rawat et al. [44] estimate the fetal weight using the femur length as shown in Figure 2; the weight of the fetus is compared with a gold standard, and then the abnormality in the fetus is predicted.

Figure 2: (a) Original 24-week femur region image. (b) Femur region superimposed onto the original image.

The length of the femur is also measured by applying the morphological operator; by this approach, automatically the FL of the fetus is measured. They projected two methods to extract the femur bone of the fetus: one is based on the entropy approach and another on edge detection. The entropy-based method is the main approach, and when the first one is failed, then the second method was only used.

3.1.4. Gradient Vector Flow (GVF) Methodology

The GVF snake is a segmentation approach [45] which has been effectively used in the segmentation of medical images. The contour of a snake [46] does not converge to the object boundary. In the image domain, the contour is initialized by the operator and then the boundary is formed in an object. According to the differential equation of GVF, the modified form of the elastic contour is defined as an external force. The vector field of the two-dimensional function is which minimizes the following objective function: where is the energy function, are the field derivatives, is the regularization parameter, and is the gradient of the edge map. Chalana et al. [4749] report an active contour model for segmentation of the fetal head and abdomen in the US images. In the physical correction in the image, it can get trapped in the local minima. Also, due to the texture inside the fetal head, the algorithm does not make the model which means that the appearance information is not used to change the accuracy. Jardim and Figueiredo [50] report the parametric deformable shape methodology for the segmentation of fetal parameters. A weakness of this method is that the optimal solution of the problem does not assure the observation of the authors. Another drawback is that the Rayleigh distribution-based model cannot take into account the spatial structure of textural patterns. The wavelet-based techniques [51] and iterative Hough transforms [52] are also useful in extracting the object or segmenting the fetal images.

In 2008, the abdominal circumference is measured by the fuzzy and gradient vector flow (GVF) methods. In the GVF method, an active contour is formed and the GVF field behaves as an external force. After applying the above method, the fetal weight is estimated using the abdominal circumference; the comparison between the accuracy of the automatic and manual measurements were presented [53]. Further, a GVF snake is reported by Nithya and Madheswaran [54] to form or extract the contour of the abdominal circumference. The value of the abdominal circumference is used to detect the intrauterine growth retarded (IUGR) fetus. The IUGR fetus is at higher risk for slow development, abdomen problem, cardiac disease, and other problems in adult life. Yang et al. [55] detected the fetal head region using the Hough transform-based classifier. In this work, a quadratic polynomial model used to assess the HC using least square fitting methods is defined as where indicates the gestational age, and and are the coefficients. Further, the manual and automatic results are compared and it is concluded that the difference between them is not considerable.

In 2014, authors applied various segmentation method for assesssment of fetal femur, fetal head and abdomen. They evaluate the results on the basis of region-based metrics which is verified by various experts [13]. Ponomarev et al. [14] applied resulting binary images with combined numerous thresholds, edge detection, and shape-based recognition. The gestational sac diameter has been used as the first fetal parameter for confirmation of pregnancy. Chakkarwar et al. [15] worked for finding the diameter of G.Sac. In this work, two steps are followed: in the first step, the global thresholding technique was used [16], then in the second step the diameter of G.Sac was measured. Rawat et al. [17] proposed the GVF methodology for finding the G.Sac contour and measuring the diameter of G.Sac. Then, the G. Sac region of the fetus is automatically segmented from the whole image and the G.Sac diameter has been measured as shown in Figure 3.

Figure 3: (a) Original 6-week and 4-day gestational sac image. (b) G.Sac contour formed using gradient vector flow (GVF) snake.
3.1.5. Graph-Based Approach

The graph-based method is proposed to extract the head of the fetus by a semisupervised patch-based approach [56]. Many segmentation problems are solved by a fast minimization format and a nonstop min-cut divider [57] in the graph. In this method, an initial label has to be defined on every image since the method is semisupervised. Fetal BPD and OFD are measured by a graph-based approach called the circular shortest path (CSP) which is a fast automatic approach [58]. Authors have done the qualitative and quantitative analysis of the segmented results which have been verified by experts.

3.2. Nuchal Translucency

Nuchal translucency (NT) of the fetus is also an important parameter for the diagnosis and assessment of fetuses. The fluid accumulation in the nuchal region at the first trimester of the fetus is the NT thickness [59]. Down syndrome in the fetus is detected by NT thickness, so large NT thickness indicates an abnormal condition. Down syndrome fetus and trisomy of 13, 18, and 21 at 10–14 weeks of gestational age have 3 mm NT thickness [60]. The bigger NT thickness indicates the structural defects and genetic syndromes even in the normal karyotype fetus [61]. NT thickness in the 10–14 weeks of gestation has been evidenced to be one of the most perceptive parameters [62].

An automatic scheme is proposed by Deng et al. [63] to estimate the fetal thickness of NT, using a filtering technique. In this technique, the initial contour is first created and extracts a preliminary contour by the GVF methodology. Then, for finding the final NT contour and computing the edge map, dynamic programming is used. Finally, NT thickness and the NT area of the fetus are calculated as shown in Figure 4.

Figure 4: (a) Original 12-week fetus image; (b) abnormal NT thickness.

Nirmala et al. [64] measure the thickness of NT to recognize chromosomal abnormalities in the first trimester fetus in three steps. In the first step, the preprocessing techniques are applied for filtering the images; in the second step, mean shift analysis [65] has been done for segmenting the NT region. Next, Canny operators for edge detection have been applied and by Blob analysis the exact thickness of the NT has been predicted. All segmentation techniques which are used for segmenting the fetal parameters are described in Table 1, and comparison of all methods are described in Table 2.

Table 1: Overview of ultrasound image segmentation techniques. A listing of popular feature extraction and classification methods for fetal US.
Table 2: Comparative analysis of important fetal image segmentation techniques.

4. Future Trends Based on the Supervised Learning Method

Previously, an assortment of the segmentation algorithm such as threading [66] and edge detection [6772] techniques have been applied on ultrasound images, for extracting the fetal parameters. The segmented region has simple and spatially accurate boundaries. This accomplishes major difficulties, since ultrasound medical images have a small hole and boundaries are also irregular. The following may be the future segmentation trends, for achieving the accurate detection task from fetal ultrasound images.

4.1. Neural Network Based on the Hybrid Approach

The neural network-based approach has been generally used in the medical field for diagnosis [73]. In the diagnosis, raw data obtained from patients are evaluated and then various artificial intelligence techniques are applied for classification or detection. Chuang et al. [74] proposed the artificial neural network (ANN) model for assessment of fetal weight and concluded that the errors are less between the calculated fetal weight and the actual fetal weight. The weight of the fetus is the indication of anomaly finding in the fetus. The accurate weight of fetus measurement is a desirable task, although previously the ANN model is used for fetal analysis, which belongs to the macrosomia group [75]. Further, the ANN model [76] is designed for diagnosis of IUGR disease in the fetus. Khashman and Curtis [77, 78] proposed the neural network model for edge detection of the fetal head and abdomen automatically. Previously, the backpropagation algorithm is applied for detection of fetal anomaly based on the head and abdominal circumference [79]. In 2011, Anjit et al. [80] proposed the ANN model for extraction of the fetal parameter of the nasal bone region of US images. Nasal region parameters are extracted in the spatial domain and converted into the spatial domain by using discrete cosine transform and wavelet transform. The training of these networks consists of mapping between an input data and a set of output data. This mapping is trained by adjusting the weights by learning the algorithm followed by the generalized delta rule [81]. In the ANN model, weights are adjusted on the training set then their value is stable and the unknown input vectors are classified. According to the generalized delta rule, the error term minimization is defined as

In this equation, the index represents the input vector, and is the target vector and are the actual output vectors. where is the rate of learning, is the local gradient, and is the change in weight from node to .

In 2014, the authors presented a new hybrid approach for detection of the IUGR fetus, using the variational level set method. Level set methods [82] are applied across fetus images for measuring the BPD and head region. The BPD and head circumference values are the test data for classification problem in the neural model. An enhanced MLP network is presented for the detection and classification of the IUGR fetus [83]. The accuracy of the IUGR fetus is calculated by measuring the statistical parameter. A multilayer perceptron network with the hybrid approach is widely used in medical image segmentation [8487].

4.2. Support Vector Machine (SVM) Approach

SVM is a classification technique for a two-group categorization problem proposed by Cortes and Vapnik [88]. The SVM model separates the positive classes (+1) and the negative classes (−1) by an optimal hyper plane. The separation between the two classes is maximized by finding the linear optimal hyper plane [89].

The SVM model in an object has M training data points , where real integer and . In the SVM algorithm, the hyper plane is indicated by (w, b) where w is the weight vector and b is the bias; x is the object classifier. In the SVM model, the data is not linearly separable, then nonlinear data points are changed to the higher-dimensional space; the data points then become linearly separable.

In 2014, Qasem et al. [90] proposed the radial basis function (RBF) kernel for breast cancer mass identification in the images. For the diagnosis of breast cancer, first of all apply segmentation algorithm across breast US images. Then, the breast images are evaluated on the basis of comparison with the ground images. Each pixel in the resulting image is compared with the equivalent pixel in the ground images firstly. Then, the confusion matrix is calculated from the resulting image with and without the use of the rejection model. Further, Hassanien and Kim [91] introduce a fusion approach that associates the fuzzy logic, SVM model, pulse coupled neural networks, and wavelet-based algorithm. In the MRI images, the SVM classifier gives the result in two categories: the first is cancerous and the second is noncancerous. Comparing with other classifier SVMs gives a more accurate result.

5. Conclusions

In this paper, a segmentation evaluation of current trends for fetal parameters is briefly reviewed. The fetal parameters can give the prediction of fetal abnormality, so accurate measurement of these parameters is of prime concern. After discussions and various simulation results were obtained, we find that the shape of fetal parameters is different, so the GVF contour method is excellent for elliptical shape parameters (AC, HC, BPD, and NT region) and morphology-based techniques are good for measuring the femur length of the fetus. A graphical approach is found better for the femur and head contour measurement of the fetus. After feature extraction, the classification techniques (neural network and support vector machine) are applied in predicting the abnormalities of the fetus. The high-risk pregnancies can be detected easily by the precise monitoring of the fetus with time and is more accurate using automated segmentation techniques. Computer-based techniques are accurate, and the speed of the algorithm is also very fast. But in the case of multiple or twin pregnancy, the parameters are not detected easily and iteration time and computational time are larger in the active contour method.

Current trends are based on an advance contour algorithm for segmentation, and a neural network-based hybrid approach and support vector machine classifier may be applied for fetus abnormality prediction. In future research, the diagnosis of medical images by the segmentation process and artificial neural model will help in improving the accuracy, precision, and computational speed. The computational-based approach also reduces the manual interaction. Further research is based on early and accurate detection of fetus status at a cheaper cost. The health care system and equipment are enhanced by the advance techniques for assisting the radiologist in making decisions effectively.

Conflicts of Interest

The authors declare that they have no conflicts of interest.

Acknowledgments

The authors are thankful to Dr. Summer Raghuvansi, Assistant Professor under the Department of Radio Diagnosis, Peoples Medical College, Bhopal, for providing an expert view on the test images.

References

  1. N. Malhotra, P. Kumar, S. Dasgupta, and R. Rajan, Ultrasound in Obstetrics and Gynecology, Jaypee Federation of Obstetrics and Gynecological Societies of India, 3rd edition, 2003.
  2. M. W. Miller, A. A. Brayman, and J. S. Abramowicz, “Obstetric ultrasonography: a biophysical consideration of patient safety—the “rules” have changed,” American Journal of Obstetrics & Gynecology, vol. 179, no. 1, pp. 241–254, 1998. View at Publisher · View at Google Scholar
  3. R. Sanders and A. James, The Principles and Practice of Ultrasonography in Obstetrics and Gynecology, Appleton-century-crofts, Norwalk, CT, USA, 3rd edition, 1985.
  4. S. L. Bridal, J. M. Correas, A. Saied, and P. Laugier, “Milestones on the road to higher resolution, quantitative, and functional ultrasonic imaging,” Proceedings of the IEEE, vol. 91, no. 10, pp. 1543–1561, 2003. View at Publisher · View at Google Scholar · View at Scopus
  5. P. Radhanakrishanan, “Referral for abortion,” Indian Journal of Medical Ethics, vol. 6, no. 4, pp. 220-221, 2009. View at Publisher · View at Google Scholar
  6. F. M. McAuliffe, L. K. Hornberger, S. Winsor, D. Chitayat, K. Chong, and J.-A. Johnson, “Fetal cardiac defects and increased nuchal translucency thickness: a prospective study,” American Journal of Obstetrics & Gynecology, vol. 191, no. 4, pp. 1486–1490, 2004. View at Publisher · View at Google Scholar · View at Scopus
  7. E. Christine, W. Hanna, A. Bakr, and M. Youssef, “Automated measurements in obstetric ultrasound images,” in Proceedings of International Conference on Image Processing, pp. 504–507, Santa Barbara, CA, USA, 1997. View at Publisher · View at Google Scholar
  8. M. Pramanik, M. Gupta, and K. B. Krishnan, “Enhancing reproducibility of ultrasonic measurements by new users,” in Medical Imaging 2013: Image Perception, Observer Performance, and Technology Assessment, vol. 8673, pp. 76–83, Lake Buena Vista, FL, USA, 2013. View at Publisher · View at Google Scholar · View at Scopus
  9. G. Carneiro, B. Georgescu, and S. Good, Knowledge-Based Automated Fetal Biometrics Using Syngo Auto OB, Measurements, Siemens Medical Solutions, 2008.
  10. J. Espinoza, S. Good, E. Russell, and W. Lee, “Does the use of automated fetal biometry improve clinical work flow efficiency,” Journal of Ultrasound in Medicine, vol. 32, no. 5, pp. 847–850, 2013. View at Publisher · View at Google Scholar
  11. J. S. Duncan and N. Ayache, “Medical image analysis: progress over two decades and the challenges ahead,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. 22, no. 1, pp. 85–106, 2000. View at Publisher · View at Google Scholar · View at Scopus
  12. R. M. Haralick and L. G. Shapiro, “Image segmentation techniques,” Computer, Vision, Graphics, and Image Processing, vol. 29, no. 1, pp. 100–132, 1985. View at Publisher · View at Google Scholar
  13. S. Rueda, S. Fathima, C. L. Knight et al., “Evaluation and comparison of current fetal ultrasound image segmentation methods for biometric measurements: a grand challenge,” IEEE Transactions on Medical Imaging, vol. 33, no. 4, pp. 797–813, 2014. View at Publisher · View at Google Scholar · View at Scopus
  14. G. V. Ponomarev, M. S. Gelfand, and M. D. Kazanov, “A multilevel thresholding combined with edge detection and shape-based recognition for segmentation of fetal ultrasound images,” in Proceedings of Challenge US: Biometric Measurements from Fetal Ultrasound Images, pp. 17–19, Barcelona, Spain, 2012.
  15. V. A. Chakkarwar, M. S. Joshi, and P. S. Revankar, “Automated analysis of gestational sac in medical image processing,” in 2010 IEEE 2nd International Advance Computing Conference (IACC), pp. 304–309, Patiala, India, 2010. View at Publisher · View at Google Scholar · View at Scopus
  16. R. C. Gonzalez and R. E. Woods, Digital Image Processing, Addison-Wesley Longman Publishing Co, 1992.
  17. V. Rawat, A. Jain, V. Shrimali, and A. Rawat, “Automatic assessment of foetal biometric parameter using GVF snakes,” International Journal of Biomedical Engineering and Technology, vol. 12, no. 4, pp. 321–233, 2013. View at Publisher · View at Google Scholar · View at Scopus
  18. A. Fenster and D. B. Downey, “3-D ultrasound imaging: a review,” IEEE Engineering in Medicine and Biology Magazine, vol. 15, no. 6, pp. 41–51, 1996. View at Publisher · View at Google Scholar · View at Scopus
  19. J. Anquez, E. D. Angelini, G. Grangé, and I. Bloch, “Automatic segmentation of antenatal 3-D ultrasound images,” IEEE Transactions on Biomedical Engineering, vol. 60, no. 5, pp. 1388–1400, 2013. View at Publisher · View at Google Scholar · View at Scopus
  20. M. Odeh, Y. Hirsh, S. Degani, V. Grinin, E. Ofir, and J. Bornstein, “Three-dimensional sonographic volumetry of the gestational sac and the amniotic sac in the first trimester,” Journal of Ultrasound in Medicine, vol. 27, no. 3, pp. 373–378, 2008. View at Publisher · View at Google Scholar · View at Scopus
  21. H.-G. K. Blaas, P. Taipale, H. Torp, and S. H. Eik-Nes, “Three-dimensional ultrasound volume calculations of human embryos and young fetuses: a study on the volumetry of compound structures and its reproducibility,” Ultrasound in Obstetrics & Gynecology, vol. 27, no. 6, pp. 640–646, 2006. View at Publisher · View at Google Scholar · View at Scopus
  22. O. Falcon, C. F. A. Peralta, P. Cavoretto, S. Faiola, and K. H. Nicolaides, “Fetal trunk and head volume measured by three-dimensional ultrasound at 11 + 0 to 13 + 6 weeks of gestation in chromosomally normal pregnancies,” Ultrasound in Obstetrics & Gynecology, vol. 26, no. 3, pp. 263–266, 2005. View at Publisher · View at Google Scholar · View at Scopus
  23. C. F. A. Peralta, P. Cavoretto, B. Csapo, O. Falcon, and K. H. Nicolaides, “Lung and heart volumes by three-dimensional ultrasound in normal fetuses at 12–32 weeks’ gestation,” Ultrasound in Obstetrics & Gynecology, vol. 27, no. 2, pp. 128–133, 2006. View at Publisher · View at Google Scholar · View at Scopus
  24. N. J. Raine-Fenning, J. S. Clewes, N. R. Kendall, A. K. Bunkheila, B. K. Campbell, and I. R. Johnson, “The interobserver reliability and validity of volume calculation from three-dimensional ultrasound datasets in the in vitro setting,” Ultrasound in Obstetrics & Gynecology, vol. 21, no. 3, pp. 283–291, 2003. View at Publisher · View at Google Scholar · View at Scopus
  25. D. T. Kuan, A. A. Sawchuk, T. C. Strand, and P. Chavel, “Adaptive noise smoothing filter for images with signal-dependent noise,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. PAMI-7, no. 2, pp. 165–177, 1985. View at Publisher · View at Google Scholar
  26. C. B. Burckhardt, “Speckle in ultrasound B-mode scans,” IEEE Transactions on Sonics and Ultrasonics, vol. 25, no. 1, pp. 1–6, 1978. View at Publisher · View at Google Scholar · View at Scopus
  27. J. M. Thijssen, “Ultrasonic speckle formation, analysis and processing applied to tissue characterization,” Pattern Recognition Letters, vol. 24, no. 4-5, pp. 659–675, 2003. View at Publisher · View at Google Scholar · View at Scopus
  28. R. W. Prager, H. Gee, G. M. Treece, and L. Berman, “Speckle detection in ultrasound images using first order statistics,” Tech. Rep., Tech. Rep. GUED/F-INFENG/TR 415, University of Cambridge, Dept. of Engineering, 2002. View at Google Scholar
  29. A. F. de Araujo, C. E. Constantinou, and J. M. R. S. Tavares, “New artificial life model for image enhancement,” Expert Systems with Applications, vol. 41, no. 13, pp. 5892–5906, 2014. View at Publisher · View at Google Scholar · View at Scopus
  30. S. M. M. Roomi and R. B. J. Rajee, “Speckle noise removal in ultrasound images using particle swarm optimization technique,” in 2011 International Conference on Recent Trends in Information Technology (ICRTIT), pp. 926–931, Chennai, Tamil Nadu, 2011. View at Publisher · View at Google Scholar · View at Scopus
  31. S. Wang, J. Zhou, J. Li, and L. Jiao, “Speckle noise reduction for us images via adaptive neighborhood accumulates multiscale product thresholding,” in Proceedings of the Second Sino-foreign-Interchange Conference on Intelligent Science and Intelligent Data Engineering, IScIDE'11, pp. 397–404, Xi'an, China, October 2011.
  32. V. Rawat, A. Jain, and V. Shrimali, “Investigation and assessment of disorder of ultrasound B-mode images,” International Journal of Computer Science and Information Security, vol. 7, no. 2, pp. 1120–1135, 2010. View at Google Scholar
  33. E. P. Simoncelli and E. H. Adelson, “Noise removal via Bayesian wavelet coring,” in Proceedings of 3rd IEEE International Conference on Image Processing, vol. 1, pp. 379–382, Lausanne, Switzerland, 1996. View at Publisher · View at Google Scholar
  34. G. Carneiro, B. Georgescu, S. Good, and D. Comaniciu, “Automatic fetal measurements in ultrasound using constrained probabilistic boosting tree,” in Medical Image Computing and Computer-Assisted Intervention – MICCAI 2007, vol. 4792 of Lecture Notes in Computer Science, pp. 571–579, Springer, Berlin, Heidelberg, 2007. View at Publisher · View at Google Scholar
  35. G. Carneiro, B. Georgescu, S. Good, and D. Comaniciu, “Detection and measurement of fetal anatomies from ultrasound images using a constrained probabilistic boosting tree,” IEEE Transactions on Medical Imaging, vol. 27, no. 9, pp. 1342–1355, 2008. View at Publisher · View at Google Scholar · View at Scopus
  36. J. Yu, Y. Wang, and P. Chen, “Fetal ultrasound image segmentation system and Its use in fetal weight estimation,” Medical & Biological Engineering & Computing, vol. 46, no. 12, pp. 1227–1237, 2008. View at Publisher · View at Google Scholar · View at Scopus
  37. Z. Tu, “Probabilistic boosting-tree: learning discriminative models for classification, recognition, and clustering,” in Tenth IEEE International Conference on Computer Vision (ICCV'05) Volume 1, vol. 2, pp. 1589–1596, Beijing, China, 2005. View at Publisher · View at Google Scholar · View at Scopus
  38. B. J. Smith and P. Arabshahi, “A fuzzy decision system for ultrasonic prenatal examination enhancement,” in Proceedings of IEEE 5th International Fuzzy Systems, vol. 3, pp. 1712–1717, New Orleans, LA, USA, 1996. View at Publisher · View at Google Scholar
  39. S. V. B. Jardim and M. A. T. Figuiredo, “Automatic analysis of fetal echographic images,” in International Conference on Image Processing, ICIP 2003, pp. 1065–1068, Catalonia, Spain, 2003.
  40. B. P. Shan and M. Madheswaran, “Extraction of fetal biometrics using class separable shape sensitive approach for gestational age estimation,” in 2009 International Conference on Computer Technology and Development, pp. 376–380, Kota Kinabalu, Malaysia, 2009. View at Publisher · View at Google Scholar · View at Scopus
  41. I. Larrabide, R. A. Feijóo, A. A. Novotny, E. Taroco, and M. Masmoudi, “An image segmentation method based on a discrete version of the topological derivative,” in 6th World Congress on Structural and Multidisciplinary Optimization, pp. 745–749, Rio de Janerio, Brazil, 2005.
  42. J. G. Thomas, R. A. Peters, and P. Jeanty, “Automatic segmentation of ultrasound images using morphological operators,” IEEE Transactions on Medical Imaging, vol. 10, no. 2, pp. 180–186, 1991. View at Publisher · View at Google Scholar · View at Scopus
  43. V. Shrimali, R. S. Anand, and V. Kumar, “Improved segmentation of ultrasound images for fetal biometry, using morphological operators,” in 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society, pp. 459–462, Minneapolis, MN, USA, 2009. View at Publisher · View at Google Scholar · View at Scopus
  44. V. Rawat, A. Jain, and V. Shrimali, “Analysis and assessment of ultrasound images for fetal biometry using morphological operators,” in 5th Indian international conference on artificial intelligence (IICAI-11), pp. 1271–1279, Tumkur, India, 2011.
  45. C. Xu and J. Prince, “Gradient vector flow: a new external force for snakes,” in Proceedings of IEEE Computer Society Conference on Computer Vision and Pattern Recognition, p. 71, San Juan, PR, USA, 1997. View at Publisher · View at Google Scholar
  46. M. Kass, A. Witkin, and D. Terzopoulos, “Snakes: active contour models,” in The Proceedings of the International Conference on Computer Vision, pp. 259–268, London, UK, 1987.
  47. V. Chalana, T. C. Winter III, D. R. Cyr, D. R. Haynor, and Y. Kim, “Automatic fetal head measurements from sonographic images,” Academic Radiology, vol. 3, no. 8, pp. 628–635, 1996. View at Publisher · View at Google Scholar · View at Scopus
  48. V. Chalana and Y. Kim, “A methodology for evaluation of boundary detection algorithms on medical images,” IEEE Transactions on Medical Imaging, vol. 16, no. 5, pp. 642–652, 1997. View at Publisher · View at Google Scholar · View at Scopus
  49. S. D. Pathak, V. Chalana, and Y. Kim, “Interactive automatic fetal head measurements from ultrasound images using multimedia computer technology,” Ultrasound in Medicine and Biology, vol. 23, no. 5, pp. 665–673, 1997. View at Publisher · View at Google Scholar · View at Scopus
  50. S. M. G. V. B. Jardim and M. A. T. Figueiredo, “Segmentation of fetal ultrasound images,” Ultrasound in Medicine and Biology, vol. 31, no. 2, pp. 243–250, 2005. View at Publisher · View at Google Scholar · View at Scopus
  51. N. M. Zayed, A. M. Badwi, A. Elsayad, M. S. Elsherif, and A. B. M. Youssef, “Wavelet segmentation for fetal ultrasound images,” in Proceedings of the 44th IEEE 2001 Midwest Symposium on Circuits and Systems. MWSCAS 2001 (Cat. No.01CH37257), vol. 1, pp. 501–504, Dayton, OH, USA, 2001. View at Publisher · View at Google Scholar · View at Scopus
  52. W. Lu, J. Tan, and R. Floyd, “Automated fetal head detection and measurement in ultrasound images by iterative randomized Hough transform,” Ultrasound in Medicine and Biology, vol. 31, no. 7, pp. 929–936, 2005. View at Publisher · View at Google Scholar · View at Scopus
  53. Y. Jinhua, Y. Wang, P. Chen, and Y. Shen, “Fetal abdominal contour extraction and measurement in ultrasound images,” Ultrasound in Medicine and Biology, vol. 34, no. 2, pp. 169–182, 2008. View at Publisher · View at Google Scholar · View at Scopus
  54. J. Nithya and M. Madheswaran, “Detection of intrauterine growth retardation using fetal abdominal circumference,” in 2009 International Conference on Computer Technology and Development, pp. 371–375, Kota Kinabalu, Malaysia, 2009. View at Publisher · View at Google Scholar · View at Scopus
  55. D. Ni, Y. Yang, S. Li et al., “Learning based automatic head detection and measurement from fetal ultrasound images via prior knowledge and imaging parameters,” in 2013 IEEE 10th International Symposium on Biomedical Imaging, pp. 772–775, San Francisco, CA, USA, 2013. View at Publisher · View at Google Scholar · View at Scopus
  56. A. Ciurte, X. Bresson, and M. B. Cuadra, “A semi-supervised patch based approach for segmentation of fetal ultrasound imaging,” in Proceedings of Challenge US: Biometric Measurements from Fetal Ultrasound Images, pp. 5–7, Barcelona, Spain, 2012.
  57. N. Houhou, X. Bresson, A. Szlam, T. F. Chan, and J.-P. Thiran, “Semi supervised segmentation based on non-local continuous min-cut,” in Proceedings of the 2nd International Conference on Scale Space and Variational Methods in Computer Vision, pp. 112–123, Voss, Norway, 2009. View at Publisher · View at Google Scholar · View at Scopus
  58. C. Sun, “Automatic fetal head measurements from ultrasound images using circular shortest paths,” in Proceedings of Challenge US: Biometric Measurements from Fetal Ultrasound Images, pp. 13–15, Barcelona, Spain, 2012.
  59. M. A. Zoppi, R. M. Ibba, M. Floris, F. Manca, C. Axiana, and G. Monni, “Changes in nuchal translucency thickness in normal and abnormal karyotype fetuses,” BJOG: An International Journal of Obstetrics and Gynaecology, vol. 110, no. 6, pp. 584–588, 2003. View at Publisher · View at Google Scholar · View at Scopus
  60. N. Montenegro, A. Matias, J. C. Areias, S. Castedo, and H. Barre, “Increased fetal nuchal translucency: possible involvement of early cardiac failure,” Ultrasound in Obstetrics & Gynecology, vol. 10, no. 4, pp. 265–268, 1997. View at Publisher · View at Google Scholar · View at Scopus
  61. A. P. Souka, E. Krampl, S. Bakalis, V. Heath, and K. H. Nicolaides, “Outcome of pregnancy in chromosomally normal fetuses with increased nuchal translucency in the first trimester,” Ultrasound in Obstetrics & Gynecology, vol. 18, no. 1, pp. 9–17, 2001. View at Publisher · View at Google Scholar · View at Scopus
  62. R. J. M. Snijders, P. Noble, N. Sebire, A. Souka, and K. H. Nicolaides, “UK multicentre project on assessment of risk of trisomy 21 by maternal age and fetal nuchal-translucency thickness at 10–14 weeks of gestation,” The Lancet, vol. 352, no. 9125, pp. 343–346, 1998. View at Publisher · View at Google Scholar · View at Scopus
  63. Y. H. Deng, Y. Y. Wang, and P. Chen, “Estimating fetal nuchal translucency parameters from its ultrasound image,” in 2008 2nd International Conference on Bioinformatics and Biomedical Engineering, pp. 2665–2668, Shanghai, China, 2008. View at Publisher · View at Google Scholar · View at Scopus
  64. S. Nirmala and V. Palanisamy, “Measurement of nuchal translucency thickness for detection of chromosomal abnormalities using first trimester ultrasound fetal images,” International Journal of Computer Science and Information Security, vol. 6, no. 3, 2009. View at Google Scholar
  65. H. Zhou, X. Li, G. Schaefer, M. E. Celebi, and P. Miller, “Mean shift based gradient vector flow for image segmentation,” Computer Vision and Image Understanding, vol. 117, no. 9, pp. 1004–1016, 2013. View at Publisher · View at Google Scholar · View at Scopus
  66. Y. Zimmer, R. Tepper, and S. Akselrod, “A two-dimensional extension of minimum cross entropy thresholding for the segmentation of ultrasound images,” Ultrasound in Medicine and Biology, vol. 22, no. 9, pp. 1183–1190, 1996. View at Publisher · View at Google Scholar · View at Scopus
  67. R. M. Haralick, “Digital step edges from zero crossing of second directional derivatives,” IEEE Transactions on Pattern Analysis and Machine Intelligence, vol. PAMI-6, no. 1, pp. 58–68, 1984. View at Publisher · View at Google Scholar · View at Scopus
  68. D. Marr and E. Hildreth, “Theory of edge detection,” Proceedings of the Royal Society B: Biological Sciences, vol. 207, no. 1167, pp. 187–217, 1980. View at Publisher · View at Google Scholar · View at Scopus
  69. S. Y. Wan and W. E. Higgins, “Symmetric region growing,” IEEE Transactions on Image Processing, vol. 12, no. 9, pp. 1007–1015, 2003. View at Publisher · View at Google Scholar · View at Scopus
  70. J. A. Noble and D. Boukerroui, “Ultrasound image segmentation: a survey,” IEEE Transactions on Medical Imaging, vol. 25, no. 8, pp. 987–1010, 2006. View at Publisher · View at Google Scholar · View at Scopus
  71. V. Shrimali, R. S. Anand, and V. Kumar, “Current trends in segmentation of medical ultrasound B-mode images: a review,” IETE Journal of Research, vol. 26, no. 1, pp. 8–17, 2009. View at Google Scholar
  72. Z. Ma, J. M. R. S. Tavares, R. N. Jorge, and T. Mascarenhas, “A review of algorithms for medical image segmentation and their applications to the female pelvic cavity,” Computer Methods in Biomechanics and Biomedical Engineering, vol. 13, no. 2, pp. 235–246, 2010. View at Publisher · View at Google Scholar · View at Scopus
  73. S. S. Cross, R. F. Harrison, and R. L. Kennedy, “Introduction to neural networks,” The Lancet, vol. 346, no. 8982, pp. 1075–1079, 1995. View at Publisher · View at Google Scholar · View at Scopus
  74. L. Chuang, J.-Y. Hwang, C.-H. Chang, C.-H. Yu, and F.-M. Chang, “Ultrasound estimation of fetal weight with the use of computerized artificial neural network model,” Ultrasound in Medicine and Biology, vol. 28, no. 8, pp. 991–996, 2002. View at Publisher · View at Google Scholar · View at Scopus
  75. R. M. Farmer, A. L. Medearis, G. I. Hirata, and L. D. Platt, “The use of a neural network for the ultrasonographic estimation of fetal weight in the macrosomic fetus,” American Journal of Obstetrics & Gynecology, vol. 166, no. 5, pp. 1467–1472, 1992. View at Publisher · View at Google Scholar · View at Scopus
  76. F. Gurgen, E. Onal, and F. G. Varol, “IUGR detection by ultrasonographic examinations using neural networks,” IEEE Engineering in Medicine and Biology Magazine, vol. 16, no. 3, pp. 55–58, 1997. View at Publisher · View at Google Scholar · View at Scopus
  77. A. Khashman and K. M. Curtis, “Automatic edge detection of foetal head and abdominal circumferences using neural network arbitration,” in Industrial Electronics, 1997. ISIE '97., Proceedings of the IEEE International Symposium on, vol. 3, pp. 1191–1194, Guimaraes, Portugal, 1997. View at Publisher · View at Google Scholar
  78. A. Khashman and K. M. Curtis, “Neural networks arbitration for automatic edge detection of 3-dimensional objects,” in Proceedings of Third International Conference on Electronics, Circuits, and Systems, pp. 49–52, Rodos, Greece, 1996. View at Publisher · View at Google Scholar
  79. V. Rawat, A. Jain, and V. Shrimali, “Automatic detection of fetal abnormality using head and abdominal circumference,” in International Conference on Computational Collective Intelligence, vol. 9876 of Lecture Notes in Computer Science, pp. 525–534, Springer, Cham, Switzerland, 2016. View at Publisher · View at Google Scholar · View at Scopus
  80. T. A. Anjit and S. Rishidas, “Identification of nasal bone for the early detection of down syndrome using back propagation neural network,” in 2011 International Conference on Communications and Signal Processing, pp. 136–140, Calicut, India, 2011. View at Publisher · View at Google Scholar · View at Scopus
  81. D. E. Rumelhart and J. L. McClelland, “Parallel distributed processing; explorations in the microstructures of cognition,” Tech. Rep., MIT Press, Cambridge, MA, USA, 1986. View at Google Scholar
  82. A. V. Gadagkar and K. S. Shreedhara, “Features based IUGR diagnosis using variational level set method and classification using artificial neural networks,” in 2014 Fifth International Conference on Signal and Image Processing, pp. 303–309, Jeju Island, South Korea, 2014. View at Publisher · View at Google Scholar · View at Scopus
  83. M. Y. Choong, M. C. Seng, S. S. Yang, A. Kiring, and K. T. K. Teo, “Foetus ultrasound medical image segmentation via variational level set algorithm,” in 2012 Third International Conference on Intelligent Systems Modelling and Simulation, pp. 225–229, Kota Kinabalu, Malaysia, 2012. View at Publisher · View at Google Scholar · View at Scopus
  84. H. Masoumi, A. Behrad, M. A. Pourmina, and A. Roosta, “Automatic liver segmentation in MRI images using an iterative watershed algorithm and artificial neural network,” Biomedical Signal Processing and Control, vol. 7, no. 5, pp. 429–437, 2012. View at Publisher · View at Google Scholar · View at Scopus
  85. A. Behrad and H. Masoumi, “Automatic spleen segmentation in MRI images using a combined neural network and recursive watershed transform,” in 10th Symposium on Neural Network Applications in Electrical Engineering, pp. 63–67, Belgrade, Serbia, 2010. View at Publisher · View at Google Scholar · View at Scopus
  86. S. Osher and J. A. Sethian, “Fronts propagating with curvature-dependent speed: algorithms based on Hamilton-Jacobi formulations,” Journal of Computational Physics, vol. 79, no. 1, pp. 12–49, 1988. View at Publisher · View at Google Scholar · View at Scopus
  87. Z. Xiao, J. Shi, and Q. Chang, “Automatic image segmentation algorithm based on PCNN and fuzzy mutual information,” in 2009 Ninth IEEE International Conference on Computer and Information Technology, pp. 241–245, Xiamen, China, 2009. View at Publisher · View at Google Scholar · View at Scopus
  88. C. Cortes and V. N. Vapnik, “Support-vector networks,” Journal of Machine Learning, vol. 20, no. 3, pp. 273–297, 1995. View at Publisher · View at Google Scholar
  89. G. Schaefer, A. E. Hassanien, and J. Jiang, Computational Intelligence in Medical Imaging Techniques and Applications, CRC Press, 2008.
  90. A. Qasem, S. N. H. S. Abdullah, and S. Sahran, “Breast cancer mass localization based on machine learning,” in 2014 IEEE 10th International Colloquium on Signal Processing and its Applications, pp. 31–36, Kuala Lumpur, Malaysia, 2014.
  91. A. E. Hassanien and T. H. Kim, “Breast cancer MRI diagnosis approach using support vector machine and pulse coupled neural networks,” Journal of Applied Logic, vol. 10, no. 4, pp. 277–284, 2012. View at Publisher · View at Google Scholar · View at Scopus
  92. C.-W. Wang, H.-C. Chen, C.-W. Peng, and C.-M. Hung, “Automatic femur segmentation and length measurement from fetal ultrasound images,” in Proceedings of Challenge US: Biometric Measurements from Fetal Ultrasound Images, pp. 21–23, Barcelona, Spain, 2012.